As shown in Figure 1, Nested Marker has a recursive layered structure. There is a marker in the upper layer and four smaller markers in the lower layer. Smaller markers also have lower-layer markers nesting inside them. Every marker at every layer is designed to be identified by its interior pattern, which is composed of lower-layer markers. The largest marker visible from the camera is dynamically selected based on the distance between observer and marker. When the observer watches the marker at close range, the lowest layer marker is used for geometric registration. When the observer watches the marker from a distance, the top-layer marker is used. Since each marker can be identified by its interior pattern, the system selects the appropriate calibration parameter set for the marker. It is also possible to simultaneously utilize all the visible markers in different layers for more stable calibration. So even if a captured marker is partially occluded by other objects, the correct camera parameter can be estimated by our Nested Marker and choosing an appropriate non-occluded marker in the captured image.